TKDE · SaGIF, our work on individual fairness in graph neural networks, accepted.
AI RESEARCH / SUN YAT-SEN UNIVERSITY
Liang Chen陈亮
Associate Professor · Computer Science & Engineering
Learning intelligence.
Connecting ideas.
I study how intelligent systems learn from language, graphs, and human interactions — and how to make that learning more trustworthy.
Based in Guangzhou, ChinaPh.D., Zhejiang University
OPEN POSITIONS
Curious minds welcome. Let’s explore what AI can become.
Research interns · Master’s · Ph.D.RESEARCH DIRECTIONS
Four perspectives.
One connected research agenda.
From the structure of data to the behavior of models.
01
Language & intelligence
Large language models, spoken dialogue, and efficient adaptation.
02
Learning from connections
Graph representation learning, temporal graphs, and spiking neural networks.
03
Trustworthy by design
Fairness, robustness, and adversarial learning for reliable AI systems.
04
Personalized intelligence
Recommendation systems that learn from relationships and interactions.
SELECTED RESEARCH
Recent work, lasting questions.
01
ICML 2025
Improving Conversational Capabilities of Speech Language Models via Generative Dual-channel Spoken Dialogue Learning
Spoken dialogue · Generative learning
LANGUAGE MODELS
02ICML 2025
Measuring Diversity in Synthetic Datasets
Synthetic data · Diversity
DATA & LEARNING
03AAAI 2026
GT-SNT: A Linear-Time Transformer for Large-Scale Graphs via Spiking Node Tokenization
Graph transformers · Spiking networks
GRAPH LEARNING
LATEST UPDATES
From the research desk.
AAAI 2026 · GT-SNT accepted: a linear-time transformer for large-scale graphs.
ICML 2025 · Two papers accepted, on synthetic data diversity and speech language models.
BEYOND THE PAPERSResearch is
Research is
a collective pursuit.
I work with students and collaborators across graph learning, language models, trustworthy AI, and recommendation systems.
Meet the group chenliang6@mail.sysu.edu.cn